Re: [PATCH v5 12/15] x86: add iommu_op to enable modification of IOMMU mappings

From: Paul Durrant <Paul.Durrant@citrix.com>
To: Kevin Tian <kevin.tian@intel.com>,
	"xen-devel@lists.xenproject.org" <xen-devel@lists.xenproject.org>
Cc: Stefano Stabellini <sstabellini@kernel.org>,
	Wei Liu <wei.liu2@citrix.com>,
	Andrew Cooper <Andrew.Cooper3@citrix.com>,
	"Tim (Xen.org)" <tim@xen.org>,
	George Dunlap <George.Dunlap@citrix.com>,
	Julien Grall <julien.grall@arm.com>,
	Jan Beulich <jbeulich@suse.com>,
	Ian Jackson <Ian.Jackson@citrix.com>
Subject: Re: [PATCH v5 12/15] x86: add iommu_op to enable modification of IOMMU mappings
Date: Tue, 7 Aug 2018 09:22:08 +0000	[thread overview]
Message-ID: <105705dc439844e782267eccae9a9be6@AMSPEX02CL03.citrite.net> (raw)
In-Reply-To: <AADFC41AFE54684AB9EE6CBC0274A5D1912AEBA5@SHSMSX101.ccr.corp.intel.com>

> -----Original Message-----
> From: Tian, Kevin [mailto:kevin.tian@intel.com]
> Sent: 07 August 2018 10:20
> To: Paul Durrant <Paul.Durrant@citrix.com>; xen-devel@lists.xenproject.org
> Cc: Stefano Stabellini <sstabellini@kernel.org>; Wei Liu
> <wei.liu2@citrix.com>; Andrew Cooper <Andrew.Cooper3@citrix.com>; Tim
> (Xen.org) <tim@xen.org>; George Dunlap <George.Dunlap@citrix.com>;
> Julien Grall <julien.grall@arm.com>; Jan Beulich <jbeulich@suse.com>; Ian
> Jackson <Ian.Jackson@citrix.com>
> Subject: RE: [Xen-devel] [PATCH v5 12/15] x86: add iommu_op to enable
> modification of IOMMU mappings
> 
> > From: Paul Durrant [mailto:Paul.Durrant@citrix.com]
> > Sent: Tuesday, August 7, 2018 5:12 PM
> >
> >
> > > So here current version pvIOMMU only provides global address space,
> > > though unlike any existing IOMMU. maybe we should explicitly call out
> > > this fact in some capability field for future extension.
> > >
> >
> > Does Linux expose such a thing through the dma_ops? I didn't see one, but
> > if there is one then we can of course set it correctly according to the
> > limitations of the underying hypervisor. Thus, if Xen is modified in future to
> > support per-device mappings, this could be exposed through to the drivers
> > using DMA.
> >
> 
> not about Linux internal exposing. it's about hypervisor exposing
> the capability to pvIOMMU driver, so the latter knows whether to
> enable per-device mapping. But as you commented in another
> mail, it could be achieved through enabling sequence.
> 

Ok. Let's go with that then.

> btw could you put clarification according to our discussions in your
> next version coverletter to set the tone clear from start? I don't
> think I'll be the only one interpreting wrong goal for this work. :-)
> 

I can certainly do that. I'll put an explanation of the XenServer use-case in as a concrete example too.

  Paul

> Thanks
> Kevin
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel